Trouble starting Tune job on local machine

Ray Libraries (Data, Train, Tune, Serve)
1

Hi, I just upgraded to Ray 2.6.1 and am having trouble with the new storage rules. I have been running a tune job on my local laptop fine until this upgrade. I have updated my code in accordance with what I understand the new instructions to be for running locally only:

    run_config = RunConfig(
                    name                        = "cda0",
                    storage_path                = "~/ray_results", #required if not using remote storage
                    stop                        = stopper,
                    sync_config                 = tune.SyncConfig(syncer = None, upload_dir = None), #for single-node or shared checkpoint dir
                    checkpoint_config           = air.CheckpointConfig(
                                                    checkpoint_frequency        = chkpt_int,
                                                    checkpoint_score_attribute  = "episode_reward_mean",
                                                    num_to_keep                 = 1,
                                                    checkpoint_at_end           = True
                    )

But it gives me the following error:
File “/home/starkj/miniconda3/envs/cda0/lib/python3.10/site-packages/ray/tune/syncer.py”, line 236, in validate_upload_dir
raise ValueError(
ValueError: upload_dir enables syncing to cloud storage, but syncer=None disables syncing. Either remove the upload_dir, or set syncer to ‘auto’ or a custom syncer.

It seems I can’t win, since I have explicitly set upload_dir = None (I have also omitted it, relying on the documented default of None). How can I go about running this with just local storage?
Thank you.

2

Update: adding debug stmts to the code for syncer.py shows that its validate_upload_dir() method is being passed the storage_path variable as the update_dir, which conflicts with my specification of update_dir = None. I also modified the init() call to be
ray.init(storage = "~/ray_results/cda0")
I suspect this is part of the problem. When I remove the storage arg here, then the job begins to run, but now I get other errors back, like
WARNING syncer.py:261 – You are using remote storage, but you don’t have fsspec installed.